|
Activity Number:
|
460
- Causal Methods for Discovery, Confirmation and Mechanistic Evaluation
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 6, 2020 : 10:00 AM to 2:00 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
|
Abstract #312994
|
|
|
Title:
|
Efficient Semiparametric Estimation of Network Treatment Effects Under Partial Interference
|
|
Author(s):
|
Chan Park* and Hyunseung Kang
|
|
Companies:
|
University of Wisconsin-Madison and University of Wisconsin-Madison
|
|
Keywords:
|
Causal inference;
Double robustness;
M-estimation;
Network causal effects;
Partial interference;
Semiparametric efficiency
|
|
Abstract:
|
There has been growing interest in causal inference to study treatment effects under interference. While many estimators have been proposed, there is little work on studying efficiency-related optimality properties of these estimators. To this end, the paper presents semiparametrically efficient and doubly robust estimation of network treatment effects under interference. We focus on partial interference where study units are partitioned into non-overlapping clusters and there is interference within clusters, but not across clusters. We derive the efficient influence function and the semiparametric efficiency bound for a family of network causal effects that include the direct and the indirect/spillover effects. We also adapt M-estimation theory to interference settings and propose M-estimators which are locally efficient and doubly robust. We conclude by presenting some limited results on adaptive estimation to interference patterns, or commonly referred to as exposure mapping.
|
Authors who are presenting talks have a * after their name.
